The proposed studies are prompted by subcutaneous adipose tissue (SAT) wasting and visceral adipose tissue VAT) accumulation, plus hyperlipidemia and insulin resistance (R1) observed in HIV-infected patients receiving highly active anti-retroviral therapies. There is concern that these developments promote atherosclerosis. However, it is not clear how the body composition changes are related to the metabolic changes, which fat compartments affect RI, nor if the risks are the same in HIV and non-HIV conditions. The underlying hypothesis is that aspects of HIV infection or its treatment decrease the influence of VAT upon RI. The specific aims are, 1) To determine the effects of HIV infection upon the relationship between body fat distribution and R1; 2) To compare the effects of diet plus exercise upon fat distribution and RI in HIV-infected and uninfected women with increased VAT; 3) To compare the effects of a thiazolidinedione, rosiglitazone, upon VAT and R1 in HIV-infected and uninfected men with increased VAT. In preliminary studies, we described the phenotypic changes that occur in HIV-infected subjects, distinguished true from pseudotruncal obesity, demonstrated the influence of sex, and determined the effects of treatment with growth hormone and resistance exercise upon VAT. We also have described the effects of race, age and total fat mass upon the relationship between VAT and R1 in non-HIV infected women, assessed possible contributions of upper body SAT to RI, and made measurements of glucose disposal rate and endogenous glucose production. A total of 180 studies are planned for PPG III. Cross sectional studies will determine the effect of HIV infection upon fat distribution and RI over a wide range of VAT, while the intervention studies will be performed in subgroups of subjects with increased VAT. Studies will include composition analysis by whole body MRI, including estimation of intramyocellular lipid, and other body composition techniques, and studies of RI by euglycemic clamp technique. Safety analyses as well as other PPG III endpoints, such as quality of life, also be measured.